1. Assessment of Urban Green Cover and its Types using Threshold-based Approach.
- Author
-
Gaikadi, S. and Vasantha Kumar, S.
- Subjects
- *
CITIES & towns , *PUBLIC spaces , *REMOTE-sensing images , *RECREATION centers , *LANDSAT satellites , *GREEN infrastructure , *TUNDRAS , *CITY dwellers - Abstract
Providing adequate green cover in the form of trees, parks and gardens is a nature-based solution which every city must adopt to address the urbanization and its impacts, as it helps to reduce the temperature, maintain the ecology and biodiversity, provide recreational facilities for the city inhabitants and many more. Accurate mapping and monitoring of Urban Green Spaces (UGS) through satellite images is essential to check whether a city has adequate green cover or not. Studies on UGS have mostly used Landsat and Sentinel-2A satellite images for extracting the green covered areas in cities. Though it is advantageous to use these freely available images, however the accuracy of extraction is a major concern as the images are of >10m spatial resolution. Distinguishing the various green cover types, such as trees, shrubs, cropland, etc. is not an easy task in low and medium resolution satellite images. To overcome these drawbacks, one has to use high (1-10m) and very-high resolution (<1m) images for extracting the UGS accurately, however, purchasing such 'cm' resolution satellite images is a costly affair. In order to balance between the cost and resolution, the best and viable option is to use the Google earth (GE) imagery and the same was attempted in the present study by proposing a simple threshold-based approach to extract the UGS and its types from GE imagery. To illustrate the proposed method, 15 sites from the city of Vellore, India were taken and the latest cloud free GE imagery was then downloaded, projected and carefully checked for its range of pixel values in different types of UGS classes. The threshold values of each UGS type were then used to reclassify the original GE image and finally the accuracy assessment was performed on the reclassified UGS map. The results were promising as the overall accuracy (OA) was > 90% at 12 sites and only 3 out of 15 sites exhibits OA in the range of 75-86%. In order to assess the performance of the proposed approach at other locations, 5 cities across the globe [2 from India (Bangalore and Hyderabad) and 3 from abroad (Barcelona, Kuala Lumpur, New York)] were taken and the results revealed that the proposed method performs well in extracting the UGS and its types as the OA was found to be above 90% at all the 5 cities. The proposed approach can be used by civic authorities not only to check whether the city has adequate green cover or not but also can monitor whether UGS has increased or decreased over the years using the historical GE images. [ABSTRACT FROM AUTHOR]
- Published
- 2024
- Full Text
- View/download PDF